The Chinese financial press is saying what Alibaba's earnings calls won't: Taobao's "dual engine" of instant commerce and AI search is losing traction. JD poaches engineers, Pinduoduo competes on price at a structural level Taobao cannot match, and Douyin owns the attention economy in ways that browse-first platforms simply can't replicate. We wrote yesterday about Alibaba's 380-billion-yuan infrastructure pledge as a moat; Securities Times is making the harder read — the moat is being dug in the wrong direction. Sir John Crabstone has more today on private capital rotating out of Alibaba and into Pinduoduo, which is where the smart money is already pointing.

The structural trap is nasty: Alibaba is too expensive to beat Pinduoduo on price, too late to own the short-video funnel Douyin locked up, and betting its infrastructure spend on a cloud and AI thesis that may not generate commerce revenue for years. That might be the right long-term bet. It is a rough short-term position for every Taobao merchant watching their traffic evaporate.

We called it: the shell was worth more than the brand. Now it is confirmed in the most literal way possible — the shoe company is sold for $39 million, the GPU-as-a-service pivot is official, and "NewBird AI" is apparently what public markets wanted all along. The brand spent six years building a sustainability story and the stock barely moved. They changed two words and gained 400%. This is not really a story about Allbirds. It is a referendum on what 2026 investors will price: AI in the name, infrastructure in the thesis, no requirement that either be connected to the original business. Someone else gets to make the shoes now.

Kingpins wrapped its Amsterdam edition with a mood that matched the headline. Geopolitical turbulence, sourcing uncertainty, tariff noise with no resolution in sight. Read this alongside today's piece from Sir John Crabstone — denim tells the tariff story before any other textile — because what gets discussed at Kingpins in April shows up in retail margins by September. The brands actively shifting sourcing to Turkey, Bangladesh, and Vietnam right now are doing so because they can see the arithmetic arriving. The ones still waiting for clarity will not like the clarity when it comes.

Southeast Asian ecommerce up 16%, driven by exactly what you would expect: TikTok Shop's video-native commerce in Indonesia, Vietnam, and Thailand, plus AI-powered discovery layering into the major platforms. We covered how TikTok Shop works even for fragrance — a category where digital commerce arguably shouldn't work at all. The metric worth watching is not topline growth; it is the ratio of browse-and-search to video-first discovery. That split is shifting fast, and the brands that have only optimized for keyword search are already losing shelf position in the region's fastest-growing channel.

Topshop — now a licensing concept with residual cultural equity — ran an AI-driven shoppable catwalk, with Shark Beauty and Lookfantastic bolting live TikTok commerce on top. This is the UK fashion industry assembling the Chinese live commerce playbook from component parts: an AI-generated event, a beauty brand with hardware credibility, a retail platform with distribution. China builds these systems end-to-end natively. Britain is building one from a mood board and three brand partnerships.

Not necessarily a failure mode — cobbled-together can work if the pieces fit — but compare this to what Pinduoduo is doing with AI interactive dramas routing viewers directly to product pages and the ambition gap is visible from space. The Topshop brand still moves UK consumers in a way that matters. The question is whether AI catwalk plus live TikTok is the right mechanism, or whether it is the channel-of-the-month playing dress-up.

Two years of digital AI deployments and now the industry is asking about stores. BoF flags the in-store AI push — assisted discovery, real-time inventory surfacing, clienteling at scale. This was always going to happen; the surprise is how long it took for "AI for our website" to become "AI for our associates." Shoptalk made clear the permission slip for announcements expired. The next proof point is going to be in-store. Every retailer that installed self-checkout in 2019 and called it innovation is now trying to figure out where to bolt an AI kiosk. The smarter play — and the one the leading operators are actually pursuing — is weaving AI into associate workflows rather than pointing customers at yet another screen. Customers do not want to talk to kiosks. Customers want associates who already know things.

Michael Kors launches an AI retail assistant on its website. In April 2026, this is table stakes — exactly the phrase Thorne's CSO used about wellness chatbots, and the same logic applies across categories. You are not differentiating by deploying a chatbot. You are matching the minimum. What matters is whether the assistant is actually useful, what customer data it collects, and whether Michael Kors does anything interesting with the interaction layer. The brand has the customer base to make this count. Most brands in their position will run the press release and then do nothing with any of it. I hope Michael Kors is the exception. I will check back in six months.

Onton just raised $7.5 million, and the timing alongside the Shopify Horizon launch below is worth noting — these are two different answers to the same question about what AI-native storefronts should look like. Shopify's answer: the platform owns design, discovery, and distribution standards; merchants ride on top. Onton's answer (presumably): some brands will pay for AI-native infrastructure that does not route control back to a single platform. The more Shopify tightens the discovery layer, the more credible the off-platform pitch becomes. $7.5M is early-stage money, but this is an intentional positioning moment, not just a funding round.

Reale Actives launched, sold out immediately, and attracted the inevitable wave of creator-brand skepticism. The CEO's statement that they "prepared for a mixed response" is the detail worth examining — it means the brand built its launch strategy around the fact that a polarizing creator generates polarized audiences, and sized inventory and messaging accordingly. That is not naive; that is running the math on what Alix Earle's specific kind of fame actually produces. The existential question, as with every creator brand, is whether the formulation justifies repurchase once the launch discourse fades. See: Forta, navigating the same question on the athletic-credibility premise. Creator brands that back up hype with real product quality build businesses. The ones that don't become cautionary tales by year three.

Alana Pallister sold her house to fund 300,000 units of a stretchy tracksuit. A million units sold. Now she is at Coachella trying to capture the next viral moment. This is fashion entrepreneurship as pure risk arbitrage: identify a silhouette with viral potential, go all-in on inventory before the market confirms you are right, then use cultural events to sustain momentum between drops. Coachella as a sequel move is smarter than it sounds — it is a content engine, not just an event. Every influencer there produces material that becomes organic distribution across platforms. That is a better play than buying impressions. Whether the follow-up product earns a second wave is the only question left.

BoF runs the piece every AI-adjacent publication eventually runs: will shopping agents hollow out the traditional ecommerce funnel? The agentic commerce thesis is real — Douyin's one-sentence shopping is already live, Shopify is explicitly building for it, and agentic storefronts are not a hypothetical. But the timeline for this to reach the median Western consumer remains years, not months. What will change sooner is the data quality war: if agents are the ones making purchase decisions, structured product data matters more than SEO-optimized title tags. Machine-readable catalogs with clean attributes, complete sizing specs, and accurate availability win in the agent layer. The brands building that infrastructure now are the ones the agents will surface in 2028. The ones still gaming search keywords for human eyeballs are going to have a very rough transition.

Etailment — the trade publication of record for Germany's retail middle market — is running the agentic commerce explainer: what it is, how retailers should prepare, what the AI shift means in practice. When the German retail trade press picks up a concept, it has crossed from VC Twitter into actual procurement meetings. German retailers adopt slowly but build durably. Once they move, they move with enterprise rigor. The failure modes we covered earlier this month — role disobedience, context collapse, cart-breaking scenarios in multi-agent systems — will matter a great deal more when conservative European operators start deploying at scale and expecting SAP-grade reliability out of systems that were designed for startups.

Drapers rounds up the UK retailers expanding their physical footprint this year, and the list is real. The pattern is clear: destination retail with genuine experience, pricing authority, or community proposition is fine; undifferentiated mid-market floor space in secondary locations is not. The brands opening stores right now are broadly those that have figured out what a store is actually for — not inventory delivery, but something a browser, an app, or a one-hour courier cannot replicate. The interesting counterpoint sits in today's piece on DoorDash's push into non-food delivery: the retailers opening stores in 2026 are implicitly betting that not everything customers want can arrive in forty minutes. That is a defensible bet. It is also the only bet they have left.

Shopify Horizon is the new AI-driven design standard for Shopify storefronts, and it is the continuation of a story we told in April: Shopify locking discovery into the platform layer. Now design follows. This is Shopify making the aesthetic baseline of online retail a platform decision rather than a merchant one. The upside for merchants is genuine — faster deployment, better-looking stores, lower development cost. The downside is structural: brand differentiation through storefront design gets harder when everyone is running the same AI-generated templates. Brand identity is increasingly a function of product, data, and community. Shopify is not causing that shift. It is just making it structurally unavoidable. Read this alongside the Onton raise above — two different visions of where the merchant's future actually lives.

Amazon marketplace sellers — not Amazon's own inventory, the third-party merchants — account for 5.5% of total UK retail sales. At that penetration, Amazon is no longer a channel you optionally participate in. It is something closer to public retail infrastructure, like payment rails or postal networks. Which makes the FBA surcharge we covered earlier this month — 17 cents per unit, zero alternatives for the majority of sellers — function less like a marketplace fee increase and more like a tax. The sellers have no meaningful leverage. They can leave Amazon, but leaving means walking away from 5.5% of the market. Almost nobody does that voluntarily. Almost nobody is going to.